6. More about the Environment

6.1. Creating an Environment

In most use cases you will interact with the Environment to do your numerical simulations. The environment is your handyman for your numerical experiments, it sets up new trajectories, keeps log files and can be used to distribute your simulations onto several CPUs.

You start your simulations by creating an environment object:

>>> env = Environment(trajectory='trajectory', comment='A useful comment')

You can pass the following arguments. Note usually you only have to change very few of these because most of the time the default settings are sufficient.

  • trajectory

    The first argument trajectory can either be a string or a given trajectory object. In case of a string, a new trajectory with that name is created. You can access the new trajectory via trajectory property. If a new trajectory is created, the comment and dynamically imported classes are added to the trajectory.

  • add_time

    Whether the current time in format XXXX_XX_XX_XXhXXmXXs is added to the trajectory name if the trajectory is newly created.

  • comment

    The comment that will be added to a newly created trajectory.

  • dynamic_imports

    Only considered if a new trajectory is created.

    The argument dynamic_imports is important if you have written your own parameter or result classes, you can pass these either as class variables MyCustomParameterClass or as strings leading to the classes in your package: 'mysim.myparameters.MyCustomParameterClass'. If you have several classes, just put them in a list dynamic_imports=[MyCustomParameterClass, MyCustomResultClass]. In case you want to load a custom class from disk and the trajectory needs to know how they are built.

    It is VERY important, that every class name is UNIQUE. So you should not have two classes named 'MyCustomParameterClass' in two different python modules! The identification of the class is based only on its name and not its path in your packages.

  • wildcard_functions

    Dictionary of wildcards like $ and corresponding functions that are called upon finding such a wildcard. For example, to replace the $ aka crun wildcard, you can pass the following: wildcard_functions = {('$', 'crun'): myfunc}.

    Your wildcard function myfunc must return a unique run name as a function of a given integer run index. Moreover, your function must also return a unique dummy name for the run index being -1.

    Of course, you can define your own wildcards like wildcard_functions = {(‘$mycard’, ‘mycard’): myfunc)}. These are not required to return a unique name for each run index, but can be used to group runs into buckets by returning the same name for several run indices. Yet, all wildcard functions need to return a dummy name for the index `-1.

    You may also want to take a look at More on Wildcards.

  • automatic_storing

    If True the trajectory will be stored at the end of the simulation and single runs will be stored after their completion. Be aware of data loss if you set this to False and not manually store everything.

  • log_config

    Can be path to a logging .ini file specifying the logging configuration. For an example of such a file see Logging. Can also be a dictionary that is accepted by the built-in logging module. Set to None if you don’t want pypet to configure logging.

    If not specified, the default settings are used. Moreover, you can manually tweak the default settings without creating a new ini file. Instead of the log_config parameter, pass a log_folder, a list of logger_names and corresponding log_levels to fine grain the loggers to which the default settings apply.

    For example:

    log_folder='logs', logger_names='('pypet', 'MyCustomLogger'), log_levels=(logging.ERROR, logging.INFO)

    You can further disable multiprocess logging via setting log_multiproc=False.

  • log_stdout

    Whether the output of stdout and stderr should be recorded into the log files. Disable if only logging statement should be recorded. Note if you work with an interactive console like IPython, it is a good idea to set log_stdout=False to avoid messing up the console output.

  • report_progress

    If progress of runs and an estimate of the remaining time should be shown. Can be True or False or a triple (10, 'pypet', logging.Info) where the first number is the percentage and update step of the resulting progressbar and the second one is a corresponding logger name with which the progress should be logged. If you use ‘print’, the print statement is used instead. The third value specifies the logging level (level of logging statement not a filter) with which the progress should be logged.

    Note that the progress is based on finished runs. If you use the QUEUE wrapping in case of multiprocessing and if storing takes long, the estimate of the remaining time might not be very accurate.

  • multiproc

    multiproc specifies whether or not to use multiprocessing (take a look at Multiprocessing). Default is False.

  • ncores

    If multiproc is True, this specifies the number of processes that will be spawned to run your experiment. Note if you use 'QUEUE' mode (see below) the queue process is not included in this number and will add another extra process for storing. If you have psutil installed, you can set ncores=0 to let psutil determine the number of CPUs available.

  • use_scoop

    If python should be used in a SCOOP framework to distribute runs amond a cluster or multiple servers. If so you need to start your script via python -m scoop my_script.py. Currently, SCOOP only works with 'LOCAL' wrap_mode (see below).

  • use_pool

    If you choose multiprocessing you can specify whether you want to spawn a new process for every run or if you want a fixed pool of processes to carry out your computation.

    When to use a fixed pool of processes or when to spawn a new process for every run? Use the former if you perform many runs (50k and more) which are inexpensive in terms of memory and runtime. Be aware that everything you use must be picklable. Use the latter for fewer runs (50k and less) and which are longer lasting and more expensive runs (in terms of memory consumption). In case your operating system allows forking, your data does not need to be picklable. If you choose use_pool=False you can also make use of the cap values, see below.

  • freeze_input

    Can be set to True if the run function as well as all additional arguments are immutable. This will prevent the trajectory from getting pickled again and again. Thus, the run function, the trajectory as well as all arguments are passed to the pool or SCOOP workers at initialisation. Works also under run_map(). In this case the iterable arguments are, of course, not frozen but passed for every run.

  • timeout

    Timeout parameter in seconds passed on to SCOOP and 'NETLOCK' wrapping. Leave None for no timeout. After timeout seconds SCOOP will assume that a single run failed and skip waiting for it. Moreover, if using 'NETLOCK' wrapping, after timeout seconds a lock is automatically released and again available for other waiting processes.

  • cpu_cap

    If multiproc=True and use_pool=False you can specify a maximum CPU utilization between 0.0 (excluded) and 100.0 (included) as fraction of maximum capacity. If the current CPU usage is above the specified level (averaged across all cores), pypet will not spawn a new process and wait until activity falls below the threshold again. Note that in order to avoid dead-lock at least one process will always be running regardless of the current utilization. If the threshold is crossed a warning will be issued. The warning won’t be repeated as long as the threshold remains crossed.

    For example let us assume you chose cpu_cap=70.0, ncores=3, and currently on average 80 percent of your CPU are used. Moreover, at the moment only 2 processes are computing single runs simultaneously. Due to the usage of 80 percent of your CPU, pypet will wait until CPU usage drops below (or equal to) 70 percent again until it starts a third process to carry out another single run.

    The parameters memory_cap and swap_cap are analogous. These three thresholds are combined to determine whether a new process can be spawned. Accordingly, if only one of these thresholds is crossed, no new processes will be spawned.

    To disable the cap limits simply set all three values to 100.0.

    You need the psutil package to use this cap feature. If not installed and you choose cap values different from 100.0 a ValueError is thrown.

  • memory_cap

    Cap value of RAM usage. If more RAM than the threshold is currently in use, no new processes are spawned. Can also be a tuple (limit, memory_per_process), first value is the cap value (between 0.0 and 100.0), second one is the estimated memory per process in mega bytes (MB). If an estimate is given a new process is not started if the threshold would be crossed including the estimate.

  • swap_cap

    Analogous to cpu_cap but the swap memory is considered.

  • niceness

    If you are running on a UNIX based system or you have psutil (under Windows) installed, you can choose a niceness value to prioritize the child processes executing the single runs in case you use multiprocessing. Under Linux these usually range from 0 (highest priority) to 19 (lowest priority). For Windows values check the psutil homepage. Leave None if you don’t care about niceness. Under Linux the niceness` value is a minimum value, if the OS decides to nice your program (maybe you are running on a server) pypet does not try to decrease the niceness again.

  • wrap_mode

    If multiproc is True, specifies how storage to disk is handled via the storage service. Since PyTables HDF5 is not thread safe, the HDF5 storage service needs to be wrapped with a helper class to allow the interaction with multiple processes.

    There are a few options:

    pypet.pypetconstants.MULTIPROC_MODE_QUEUE: (‘QUEUE’)

    Another process for storing the trajectory is spawned. The sub processes running the individual single runs will add their results to a multiprocessing queue that is handled by an additional process.

    pypet.pypetconstants.MULTIPROC_MODE_LOCK: (‘LOCK’)

    Each individual process takes care about storage by itself. Before carrying out the storage, a lock is placed to prevent the other processes to store data. Allows loading of data during runs.


    Experimental mode based on a single pipe. Is faster than 'QUEUE' wrapping but data corruption may occur, does not work under Windows (since it relies on forking).


    Data is not stored during the single runs but after they completed. Storing is only performed locally in the main process.

    Note that removing data during a single run has no longer an effect on memory whatsoever, because there are references kept for all data that is supposed to be stored.


    Similar to ‘LOCK’ but locks can be shared across a network. Sharing is established by running a lock server that distributes locks to the individual processes. Can be used with SCOOP if all hosts have access to a shared home directory. Allows loading of data during runs.


    Similar to ‘QUEUE’ but data can be shared across a network. Sharing is established by running a queue server that distributes locks to the individual processes.

    If you don’t want wrapping at all use pypet.pypetconstants.MULTIPROC_MODE_NONE (‘NONE’).

    If you have no clue what I am talking about, you might want to take a look at multiprocessing in python to learn more about locks, queues and thread safety and so forth.

  • queue_maxsize

    Maximum size of the Storage Queue, in case of 'QUEUE' wrapping. 0 means infinite, -1 (default) means the educated guess of 2 * ncores.

  • port

    Port to be used by lock server in case of 'NETLOCK' wrapping. Can be a single integer as well as a tuple (7777, 9999) to specify a range of ports from which to pick a random one. Leave None for using pyzmq’s default range. In case automatic determining of the host’s IP address fails, you can also pass the full address (including the protocol and the port) of the host in the network like 'tcp://'.

  • param gc_interval

    Interval (in runs or storage operations) with which gc.collect() should be called in case of the 'LOCAL', 'QUEUE', or 'PIPE' wrapping. Leave None for never.

    In case of 'LOCAL' wrapping 1 means after every run 2 after every second run, and so on. In case of 'QUEUE' or 'PIPE'' wrapping 1 means after every store operation, 2 after every second store operation, and so on. Only calls gc.collect() in the main (if 'LOCAL' wrapping) or the queue/pipe process. If you need to garbage collect data within your single runs, you need to manually call gc.collect().

    Usually, there is no need to set this parameter since the Python garbage collection works quite nicely and schedules collection automatically.

  • clean_up_runs

    In case of single core processing, whether all results under results.runs.run_XXXXXXXX and derived_parameters.runs.run_XXXXXXXX should be removed after the completion of the run. Note in case of multiprocessing this happens anyway since the trajectory container will be destroyed after finishing of the process.

    Moreover, if set to True after post-processing run data is also cleaned up.

  • immediate_postproc

    If you use post- and multiprocessing, you can immediately start analysing the data as soon as the trajectory runs out of tasks, i.e. is fully explored but the final runs are not completed. Thus, while executing the last batch of parameter space points, you can already analyse the finished runs. This is especially helpful if you perform some sort of adaptive search within the parameter space.

    The difference to normal post-processing is that you do not have to wait until all single runs are finished, but your analysis already starts while there are still runs being executed. This can be a huge time saver especially if your simulation time differs a lot between individual runs. Accordingly, you don’t have to wait for a very long run to finish to start post-processing.

    Note that after the execution of the final run, your post-processing routine will be called again as usual.

    IMPORTANT: If you use immediate post-processing, the results that are passed to your post-processing function are not sorted by their run indices but by finishing time!

  • resumable

    Whether the environment should take special care to allow to resume or continue crashed trajectories. Default is False.

    You need to install dill to use this feature. dill will make snapshots of your simulation function as well as the passed arguments. Be aware that dill is still rather experimental!

    Assume you run experiments that take a lot of time. If during your experiments there is a power failure, you can resume your trajectory after the last single run that was still successfully stored via your storage service.

    The environment will create several .ecnt and .rcnt files in a folder that you specify (see below). Using this data you can continue crashed trajectories.

    In order to resume trajectories use resume().

    Your individual single runs must be completely independent of one another to allow continuing to work. Thus, they should not be based on shared data that is manipulated during runtime (like a multiprocessing manager list) in the positional and keyword arguments passed to the run function.

    If you use postprocessing, the expansion of trajectories and continuing of trajectories is not supported properly. There is no guarantee that both work together.

  • resume_folder

    The folder where the resume files will be placed. Note that pypet will create a sub-folder with the name of the environment.

  • delete_resume

    If true, pypet will delete the resume files after a successful simulation.

  • storage_service

    Pass a given storage service or a class constructor (default is HDF5StorageService) if you want the environment to create the service for you. The environment will pass additional keyword arguments you provide directly to the constructor. If the trajectory already has a service attached, the one from the trajectory will be used. For the additional keyword arguments, see below.

  • git_repository

    If your code base is under git version control you can specify the path (relative or absolute) to the folder containing the .git directory. See also Git Integration.

  • git_message

    Message passed onto git command.

  • git_fail

    If True the program fails instead of triggering a commit if there are not committed changes found in the code base. In such a case a GitDiffError is raised.

  • do_single_runs

    Whether you intend to actually to compute single runs with the trajectory. If you do not intend to carry out single runs (probably because you loaded an old trajectory for data analysis), than set to False and the environment won’t add config information like number of processors to the trajectory.

  • graceful_exit

    If True hitting CTRL+C (i.e.sending SIGINT) will not terminate the program immediately. Instead, active single runs will be finished and stored before shutdown. Hitting CTRL+C twice will raise a KeyboardInterrupt as usual.

  • lazy_debug

    If lazy_debug=True and in case you debug your code (aka you use pydevd and the expression 'pydevd' in sys.modules is True), the environment will use the LazyStorageService instead of the HDF5 one. Accordingly, no files are created and your trajectory and results are not saved. This allows faster debugging and prevents pypet from blowing up your hard drive with trajectories that you probably not want to use anyway since you just debug your code.

If you use the standard HDF5StorageService you can pass the following additional keyword arguments to the environment. These are handed over to the service:

  • filename

    The name of the hdf5 file. If none is specified, the default ./hdf5/the_name_of_your_trajectory.hdf5 is chosen. If filename contains only a path like filename='./myfolder/', it is changed to filename='./myfolder/the_name_of_your_trajectory.hdf5'.

  • file_title

    Title of the hdf5 file (only important if file is created new)

  • overwrite_file

    If the file already exists it will be overwritten. Otherwise the trajectory will simply be added to the file and already existing trajectories are not deleted.

  • encoding

    Encoding for unicode characters. The default 'utf8' is highly recommended.

  • complevel

    You can specify your compression level. 0 means no compression and 9 is the highest compression level. By default the level is set to 9 to reduce the size of the resulting HDF5 file. See PyTables Compression for a detailed explanation.

  • complib

    The library used for compression. Choose between zlib, blosc, and lzo. Note that ‘blosc’ and ‘lzo’ are usually faster than ‘zlib’ but it may be the case that you can no longer open your hdf5 files with third-party applications that do not rely on PyTables.

  • shuffle

    Whether or not to use the shuffle filters in the HDF5 library. This normally improves the compression ratio.

  • fletcher32

    Whether or not to use the Fletcher32 filter in the HDF5 library. This is used to add a checksum on hdf5 data.

  • pandas_format

    How to store pandas data frames. Either in ‘fixed’ (‘f’) or ‘table’ (‘t’) format. Fixed format allows fast reading and writing but disables querying the hdf5 data and appending to the store (with other 3rd party software other than pypet).

  • purge_duplicate_comments

    If you add a result via f_add_result() or a derived parameter f_add_derived_parameter() and you set a comment, normally that comment would be attached to each and every instance. This can produce a lot of unnecessary overhead if the comment is the same for every result over all runs. If hdf5.purge_duplicate_comments=True than only the comment of the first result or derived parameter instance created is stored, or comments that differ from this first comment. You might want to take a look at HDF5 Purging of Duplicate Comments.

  • summary_tables

    Whether summary tables should be created. These give overview about ‘derived_parameters_runs_summary’, and ‘results_runs_summary’. They give an example about your results by listing the very first computed result. If you want to purge_duplicate_comments you will need the summary_tables. You might want to check out HDF5 Overview Tables.

  • small_overview_tables

    Whether the small overview tables should be created. Small tables are giving overview about ‘config’, ‘parameters’, ‘derived_parameters_trajectory’, ‘results_trajectory’.

  • large_overview_tables

    Whether to add large overview tables. These encompass information about every derived parameter and result and the explored parameters in every single run. If you want small HDF5 files set to False (default).

  • results_per_run

    Expected results you store per run. If you give a good/correct estimate, storage to HDF5 file is much faster in case you want large_overview_tables.

    Default is 0, i.e. the number of results is not estimated!

  • derived_parameters_per_run

    Analogous to the above.

Finally, you can also pass properties of the trajectory, like v_auto_load=True (you can leave the prefix v_, i.e. auto_load works, too). Thus, you can change the settings of the trajectory immediately.

6.1.1. Config Data added by the Environment

The Environment will automatically add some config settings to your trajectory. Thus, you can always look up how your trajectory was run. This encompasses many of the above named parameters as well as some information about the environment. This additional information includes a timestamp and a SHA-1 hash code that uniquely identifies your environment. If you use git integration (Git Integration), the SHA-1 hash code will be the one from your git commit. Otherwise the code will be calculated from the trajectory name, the current time, and your current pypet version.

The environment will be named environment_XXXXXXX_XXXX_XX_XX_XXhXXmXXs. The first seven X are the first seven characters of the SHA-1 hash code followed by a human readable timestamp.

All information about the environment can be found in your trajectory under config.environment.environment_XXXXXXX_XXXX_XX_XX_XXhXXmXXs. Your trajectory could potentially be run by several environments due to merging or extending an existing trajectory. Thus, you will be able to track how your trajectory was built over time.

6.2. Logging

pypet comes with a full fledged logging environment.

Per default the environment will created loggers and stores all logged messages to log files. This can include also everything written to the standard stream stdout, like print statements, for instance. In order to log print statements set log_stdout=True. log_stdout can also be a tuple: ('mylogger', 10), specifying a logger name as well as a log-level. The log-level defines with what level stdout is logged, it is not a filter.

Note that you should always disable this feature in case you use an interactive console like IPython. Otherwise your console output will be garbled.

After your experiments are finished you can disable logging to files via disable_logging(). This also restores the standard stream.

You can tweak the standard logging settings via passing the following arguments to the environment. log_folder specifies a folder where all log-files are stored. logger_names is a list of logger names to which the standard settings apply. log_levels is a list of levels with which the specified loggers should be logged. You can further disable multiprocess logging via setting log_multiproc=False.

import logging
from pypet import Environment

env =  Environment(trajectory='mytraj',
                 log_folder = './logs/',
                 logger_nmes = ('pypet', 'MyCustomLogger'),
                 log_levels=(logging.ERROR, logging.INFO),

Furthermore, if the standard settings don’t suite you at all, you can fine grain logging via a logging config file passed via log_config='/test/ini.'. This file has to follow the logging configurations of the logging module.

Additionally, if you create file handlers you can use the following wildcards in the filenames which are replaced during runtime:

LOG_ENV ($env) is replaces by the name of the trajectory`s environment.

LOG_TRAJ ($traj) is replaced by the name of the trajectory.

LOG_RUN ($run) is replaced by the name of the current run.

LOG_SET ($set) is replaced by the name of the current run set.

LOG_PROC ($proc) is replaced by the name fo the current process and its process id.

LOG_HOST ($host) is replaced by the network name of the current host (note that dots (.) in the hostname are replaced by minus (-))

Note that in contrast to the standard logging package, pypet will automatically create folders for your log-files if these don’t exist.

You can further specify settings for multiprocessing logging which will overwrite your current settings within each new process. To specify settings only used for multiprocessing, simply append multiproc_ to the sections of the .ini file.

An example logging ini file including multiprocessing is given below.

Download: default.ini




format=%(asctime)s %(name)s %(levelname)-8s %(message)s

format=%(processName)-10s %(name)s %(levelname)-8s %(message)s

keys=file_main, file_error, stream







format=%(asctime)s %(name)s %(levelname)-8s %(message)s




Furthermore, an environment can also be used as a context manager such that logging is automatically disabled in the end:

import logging
from pypet import Environment

with Environment(trajectory='mytraj',
                 log_stdout=True) as env:
    traj = env.trajectory

    # do your complex experiment...

This is equivalent to:

import logging
from pypet import Environment

env = Environment(trajectory='mytraj',
traj = env.trajectory

# do your complex experiment...


6.3. Multiprocessing

For an example on multiprocessing see Multiprocessing.

The following code snippet shows how to enable multiprocessing with 4 CPUs, a pool, and a queue.

env = Environment(self, trajectory='trajectory',

Setting use_pool=True will create a pool of ncores worker processes which perform your simulation runs.

IMPORTANT: Python multiprocessing does not work well with multi-threading of openBLAS. If your simulation relies on openBLAS, you need to make sure that multi-threading is disabled. For disabling set the environment variables OPENBLAS_NUM_THREADS=1 and OMP_NUM_THREADS=1 before starting python and using pypet. For instance, numpy and matplotlib (!) use openBLAS to solve linear algebra operations. If your simulation relies on these packages, make sure the environment variables are changed appropriately. Otherwise your program might crash or get stuck in an infinite loop.

IMPORTANT: In order to allow multiprocessing with a pool (or in general under Windows), all your data and objects of your simulation need to be serialized with pickle. But don’t worry, most of the python stuff you use is automatically picklable.

If you come across the situation that your data cannot be pickled (which is the case for some BRIAN networks, for example), don’t worry either. Set use_pool=False (and also continuable=False) and for every simulation run pypet will spawn an entirely new subprocess. The data is than passed to the subprocess by forking on OS level and not by pickling. However, this only works under Linux. If you use Windows and choose use_pool=False you still need to rely on pickle because Windows does not support forking of python processes.

Besides, as a general rule of thumb when to use use_pool or don’t: Use the former if you perform many runs (50k and more) which are in terms of memory and runtime inexpensive. Use no pool (use_pool=False) for fewer runs (50k and less) and which are longer lasting and more expensive runs (in terms of memory consumption). In case your operating system allows forking, your data does not need to be picklable. Furthermore, if your trajectory contains many parameters and you want to avoid that your trajectory gets pickled over and over again you can set freeze_input=True. The trajectory, the run function as well as the all additional function arguments are passed to the multiprocessing pool at initialization. Be aware that the run function as well as the the additional arguments must be immutable, otherwise your individual runs are no longer independent. In case you use run_map() (see below), additional arguments are not frozen but passed for every run.

Moreover, if you enable multiprocessing and disable pool usage, besides the maximum number of utilized processors ncores, you can specify usage cap levels with cpu_cap, memory_cap, and swap_cap as fractions of the maximum capacity. Values must be chosen larger than 0.0 and smaller or equal to 100.0. If any of these thresholds is crossed no new processes will be started by pypet. For instance, if you want to use 3 cores aka ncores=3 and set a memory cap of memory_cap=90. and let’s assume that currently only 2 processes are started with currently 95 percent of you RAM are occupied. Accordingly, pypet will not start the third process until RAM usage drops again below (or equal to) 90 percent.

In addition, (only) the memory_cap argument can alternatively be a tuple with two entries: (cap, memory_per_process). First entry is the cap value between 0.0 and 100.0 and the second one is the estimated memory per process in mega-bytes (MB). If you specify such an estimate, starting a new process is suspended if the threshold would be reached including the estimated memory.

Moreover, to prevent dead-lock pypet will regardless of the cap values always start at least one process. To disable the cap levels, simply set all three to 100.0 (which is default, anyway). pypet does not check if the processes themselves obey the cap limit. Thus, if one of the process that computes your single runs needs more RAM/Swap or CPU power than the cap value, this is its very own problem. The process will not be terminated by pypet. The process will only cause pypet to not start new processes until the utilization falls below the threshold again. In order to use this cap feature, you need the psutil package.

In addition to the cap values, you can also choose the niceness of your multiprocessing processes. If your operating system supports nice (Linux, MacOS) natively, this feature works even without the psutil package. Priority values under Linux usually range from 0 (highest) to 19 (lowest), for Windows values see the psutil documentation. Low priority processes will be given less CPU time, so they are nice to other processes. Nicing works with use_pool as well. Leave None if you don’t care about niceness.

Note that HDF5 is not thread safe, so you cannot use the standard HDF5 storage service out of the box. However, if you want multiprocessing, the environment will automatically provide wrapper classes for the HDF5 storage service to allow safe data storage. There are a couple different modes that are supported. You can choose between them via setting wrap_mode. You can select between 'QUEUE', 'LOCK', 'PIPE', 'LOCAL',``’NETLOCK’, and ``'NETQUEUE' wrapping. If you have your own service that is already thread safe you can also choose 'NONE' to skip wrapping.

If you chose the 'QUEUE' mode, there will be an additional process spawned that is the only one writing to the HDF5 file. Everything that is supposed to be stored is send over a queue to the process. This has the advantage that your worker processes are only busy with your simulation and are not bothered with writing data to a file. More important, they don’t spend time waiting for other processes to release a thread lock to allow file writing. The disadvantages are that you can only store but not load data and storage relies a lot on pickling of data, so often your entire trajectory is send over the queue. Moreover, in case of 'QUEUE' wrapping you can choose the queue_maxsize of elements that can be put on the queue. To few means that your worker processes may need to wait until they can put more data on the queue. To many could blow up your memory in cases the single runs are actually faster than the storage of the data. 0 means a queue of infinite size. Default is -1 meaning pypet makes a conservative estimate of twice te number of processes (i.e. 2 * ncores). This doesn’t sound a lot. However, keep in mind that a single element on the queue might already be quite large like the entire data gathered in a single run.

If you chose the 'LOCK' mode, every process will place a lock before it opens the HDF5 file for writing data. Thus, only one process at a time stores data. The advantages are the possibility to load data and that your data does not need to be send over a queue over and over again. Yet, your simulations may take longer since processes have to wait often for each other to release locks.

'PIPE' wrapping is a rather experimental mode where all processes feed their data into a shared multiprocessing pipe. This can be much faster than a queue. However, no data integrity checks are made. So there’s no guarantee that all you data is really saved. Use this if you go for many runs that just produce small results, and use it carefully. Since this mode relies on forking of processes, it cannot be used under Windows.

'LOCAL' wrapping means that all data is kept and feed back to your local main process as soon as a single run is completed. Your data is then stored by your main process. This wrap mode can be useful if you use pypet with SCOOP (see also Multiprocessing with a Cluster or a Multi-Server Framework) in a cluster environment and your workers are distributed over a network. Note that freeing data with f_empty() during a single run has no effect on your memory because the local wrapper will keep references to all data until the single run is completed.

'NETLOCK' wrapping is similar to 'LOCK' wrapping but locks can be shared across a computer network. Lock distribution is established by a server process that listens at a particular port for lock requests. The server locks and releases locks accordingly. Like regular 'LOCK' wrapping it allows to load data during the runs. This wrap mode can be used with SCOOP if all hosts have access to a shared home directory. 'NETLOCK' wrapping requires an installation of pyzmq. However, installing SCOOP will automatically install pyzmq if it is missing.

'NETQUEUE' wrapping is similar to 'QUEUE' wrapping but data can be shared across a computer network. Data is collected by a server process that listens at a particular port. As above this wrap mode can be used with SCOOP and requires pyzmq.

Finally, there also exists a lightweight multiprocessing environment MultiprocContext. It allows to use trajectories in a multiprocess safe setting without the need of a full Environment. For instance, you might use this if you also want to analyse the trajectory with multiprocessing. You can find an example here: Lightweight Multiprocessing.

6.3.1. Multiprocessing with a Cluster or a Multi-Server Framework

pypet can be used on computing clusters as well as multiple servers sharing a home directory via SCOOP.

Simply create your environment as follows

env = Environment(multiproc=True,

and start your script via python -m scoop my_script.py. If using SCOOP, the only multiprocessing wrap modes currently supported are 'LOCAL', 'NETLOCK', and 'NETQUEUE'. That is in the former case all your data is actually stored by your local main python process and results are collected from all workers. 'NETLOCK' means locks are shared across the computer network to allow only one process to write data at a time. Lastly, 'NETQUEUE' starts queue process that collects data stores it.

In case SCOOP is configured correctly, you can easily use pypet in a multi-server or cluster framework. Using SCOOP multiprocessing shows how to combine pypet and SCOOP. For instance, if you have multiple servers sharing the same home directory you can distribute your runs on all of them via python -m scoop --hostfile hosts -vv -n 16 my_script.py to start 16 workers on your hosts which is a file specifying the servers to use. It has the format

some_host 10
another_host 4

with the name or IP of the host followed by the number of workers you want to launch (optional).

To use pypet and SCOOP on a computing cluster one additional needs a bash start-up script. For instance, for a sun grid engine (SGE), the bash script might look like the following:

#$ -l h_rt=3600
#$ -N mysimulation
#$ -pe mp 4
#$ -cwd

# Launch the simulation with SCOOP
python -m scoop -vv mysimulation.py

Most important is the -pe parallel environment flag to let the computer grid and SCOOP know how many workers to spawn (here 4). Other options may be parameters like -l h_rt defining the maximum runtime, -N assigning a name, or -cwd using the the current folder as the working directory. The particular options depend on your cluster environment and requirements of the grid provider. This job script, let’s name it mybash.sh, can be submitted via

$ qsub mybash.sh

Accordingly, the simulation mysimulation.py gets queued and eventually executed in parallel on the computer grid as soon as resources are available.

See also the SCOOP docs and the example start up scripts on how to set up multiple hosts and scripts for other grid engines.

To avoid overhead of re-pickling the trajectory, SCOOP mode also supports setting freeze_input=True (see Multiprocessing).

Moreover, you can also use pypet with SAGA Python to manually schedule your experiments on a cluster environment. Using pypet with SAGA-Python shows how to submit batches of experiments and later on merge the trajectories from each experiment into one.

6.4. Git Integration

The environment can make use of version control. If you manage your code with git, you can trigger automatic commits with the environment to get a proper snapshot of the code you actually use. This ensures that your experiments are repeatable. In order to use the feature of git integration, you additionally need GitPython.

To trigger an automatic commit simply pass the arguments git_repository and git_message to the Environment constructor. git_repository specifies the path to the folder containing the .git directory. git_message is optional and adds the corresponding message to the commit. Note that the message will always be augmented with some short information about the trajectory you are running. The commit SHA-1 hash and some other information about the commit will be added to the config subtree of your trajectory, so you can easily recall that commit from git later on.

The automatic commit functionality will only commit changes in files that are currently tracked by your git repository, it will not add new files. So make sure to put new files into your repository before running an experiment. Moreover, a commit will only be triggered if your working copy contains changes. If there are no changes detected, information about the previous commit will be added to the trajectory. By the way, the autocommit function is similar to calling $ git add -u and $ git commit -m 'Some Message' in your console.

If you want git version control but no automatic commits of your code base in case of changes, you can pass the option git_fail=True to the environment. Instead of triggering a new commit in case of changed code, the program will throw a GitDiffError.

6.5. Sumatra Integration

The environment can make use of a Sumatra experimental lab-book.

Just pass the argument sumatra_project - which should specify the path to your root sumatra folder - to the Environment constructor. You can additionally pass a sumatra_reason, a string describing the reason for you sumatra simulation. pypet will automatically add the name, comment, and the names of all explored parameters to the reason. You can also pick a sumatra_label, set this to None if you want Sumatra to pick a label for you. Moreover, pypet automatically adds all parameters to the sumatra record. The explored parameters are added with their full range instead of the default values.

In contrast to the automatic git commits (see above), which are done as soon as the environment is created, a sumatra record is only created and stored if you actually perform single runs. Hence, records are stored if you use one of following three functions: run(), or pipeline(), or resume() and your simulation succeeds and does not crash.

6.6. HDF5 Overview Tables

The HDF5StorageService creates summarizing information about your trajectory that can be found in the overview group within your HDF5 file. These overview tables give you a nice summary about all parameters and results you needed and computed during your simulations.

The following tables are created depending of your choice of large_overview_tables and small_overview_tables:

  • An info table listing general information about your trajectory (needed internally)

  • A runs table summarizing the single runs (needed internally)

  • An explorations table listing only the names of explored parameters (needed internally)

  • The branch tables:


    Containing all parameters, and some information about comments, length etc.


    As above, but config parameters


    All results to reduce memory size only a short value summary and the name is given. Per default this table is switched off, to enable it pass large_overview_tables=True to your environment.


    Only the very first result with a particular comment is listed. For instance, if you create the result ‘my_result’ in all with the comment 'Contains my important data'. Only the very first result having this comment is put into the summary table.

    If you use this table, you can purge duplicate comments, see HDF5 Purging of Duplicate Comments.



    Both are analogous to the result overviews above

  • The explored_parameters_overview overview table showing the explored parameter ranges

IMPORTRANT: Be aware that overview and summary tables are only for eye-balling of data. You should never rely on data in these tables because it might be truncated or outdated. Moreover, the size of these tables is restricted to 1000 entries. If you add more parameters or results, these are no longer listed in the overview tables. Finally, deleting or merging information does not affect the overview tables. Thus, deleted data remains in the table and is not removed. Again, the overview tables are unreliable and their only purpose is to provide a quick glance at your data for eye-balling.

6.6.1. HDF5 Purging of Duplicate Comments

Adding a result with the same comment in every single run, may create a lot of overhead. Since the very same comment would be stored in every node in the HDF5 file. To get rid of this overhead use the option purge_duplicate_comments=True and summary_tables=True.

For instance, during a single run you call traj.f_add_result('my_result', 42, comment='Mostly harmless!') and the result will be renamed to results.runs.run_00000000.my_result. After storage of the result into your HDF5 file, you will find the comment 'Mostly harmless!' in the corresponding HDF5 group node. If you call traj.f_add_result('my_result',-55, comment='Mostly harmless!') in another run again, let’s say run_00000001, the name will be mapped to results.runs.run_00000001.my_result. But this time the comment will not be saved to disk, since 'Mostly harmless!' is already part of the very first result with the name ‘my_result’.

Furthermore, if you reload your data from the example above, the result instance results.runs.run_00000001.my_result won’t have a comment only the instance results.runs.run_00000000.my_result.

IMPORTANT: If you use multiprocessing, the comment of the first result that was stored is used. Since runs are performed synchronously there is no guarantee that the comment of the result with the lowest run index is kept.

IMPORTANT Purging of duplicate comments requires overview tables. Since there are no overview tables for group nodes, this feature does not work for comments in group nodes. So try to avoid to adding the same comments over and over again in group nodes within single runs.

6.7. Using a Config File

You are not limited to specify the logging environment within an .ini file. You can actually specify all settings of the environment and already add some basic parameters or config data yourself. Simply pass config='my_config_file.ini to the environment. If your .ini file encompasses logging settings, you don’t have to pass another log_config.

Anything found in an environment, trajectory or storage_service section is directly passed to the environment constructor. Yet, you can still specify other setting of the environment. Settings passed to the constructor directly take precedence over settings specified in the ini file.

Anything found under parameters or config is added to the trajectory as parameter or config data.

An example ini file including logging can be found below.

Download: environment_config.ini

######### Environment ##############

log_stdout=('STDOUT', 50)
report_progress = (10, 'pypet', 50)


###### Config and Parameters ######
test.testconfig=True, 'This is a test config'

y=43, 'This is the second variable'

############ Logging  ###############



format=%(asctime)s %(name)s %(levelname)-8s %(message)s

format=%(processName)-10s %(name)s %(levelname)-8s %(message)s

keys=file_main, file_error, stream







format=%(asctime)s %(name)s %(levelname)-8s %(message)s

keys=file_main, file_error



Example usage:

env = Environment(config='path/to/my_config.ini',
                  multiproc = False # This will set multiproc to `False` regardless of the
                  # setting within the `my_config.ini` file.

6.8. Running an Experiment

In order to run an experiment, you need to define a job or a top level function that specifies your simulation. This function gets as first positional argument the: Trajectory container (see More on Trajectories), and optionally other positional and keyword arguments of your choice.

def myjobfunc(traj, *args, **kwargs)
    # Do some sophisticated simulations with your trajectory
    return 'fortytwo'

In order to run this simulation, you need to hand over the function to the environment. You can also specify the additional arguments and keyword arguments using run():

env.run(myjobfunc, *args, **kwargs)

The argument list args and keyword dictionary kwargs are directly handed over to the myjobfunc during runtime.

The run() will return a list of tuples. Whereas the first tuple entry is the index of the corresponding run and the second entry of the tuple is the result returned by your run function. For the example above this would simply always be the string 'fortytwo', i.e. ((0, 'fortytwo'), (1, 'fortytwo'),...). These will always be in order of the run indices even in case of multiprocessing. The only exception to this rule is if you use immediate postprocessing (see Adding Post-Processing) where results are in order of finishing time.

using run() all args and kwargs are supposed to be static, that is all of them are passed to every function call. If you need to pass different values to each function call of your job function use run_map(), where each entry in args and kwargs needs to be an iterable (list, tuple, iterator, generator etc.). Hence, the contents of each iterable are passed one after the other to your job function. For instance, assuming besides the trajectory your job function takes 3 arguments (here passed as 2 positional and 1 keyword argument):

def myjobfunc(traj, arg1, arg2, arg3):
    # do stuff


env.run(myjobfunc, range(5), ['a','b','c','d','e'], arg3=[5,4,3,2,1])

Thus, the first run of your job function will be started with the arguments 0 (from range) 'a' (from the list) and arg3=5 (from the other list). Accordingly the second run gets passed 1, 'b', arg3=4.

6.8.1. Graceful Exit

Sometimes you might need to stop your experiments via CTRL-C. If you did choose graceful_exit=True when creating an environment, CTRL-C won’t kill your program immediately but pypet will try to exit gracefully. That is pypet will finish the currently active runs and wait until their results have been returned. Hitting CTRL+C twice will, of course, immediately kill your program.

By default graceful_exit is False because it does not work in all python contexts. For instance, graceful_exit does not work with IPython notebooks. If in doubt, just try it out.

6.9. Adding Post-Processing

You can add a post-processing function that is called after the execution of all the single runs via add_postprocessing().

Your post processing function must accept the trajectory container as the first argument, a list of tuples (containing the run indices and results, normally in order of indices unless you use immediate_postproc, see below), and arbitrary positional and keyword arguments. In order to pass arbitrary arguments to your post-processing function, simply pass these first to add_postprocessing().

For example:

def mypostprocfunc(traj, result_list, extra_arg1, extra_arg2):
    # do some postprocessing here

Whereas in your main script you can call

env.add_postproc(mypostprocfunc, 42, extra_arg2=42.5)

which will later on pass 42 as extra_arg1 and 42.4 as extra_arg2. It is the very same principle as before for your run function. The post-processing function will be called after the completion of all single runs.

Moreover, please note that your trajectory usually does not contain the data computed during the single runs, since this has been removed after the single runs to save RAM. If your post-processing needs access to this data, you can simply load it via one of the many loading functions (f_load_child(), f_load_item(), f_load()) or even turn on Automatic Loading.

Note that your post-processing function should not return any results, since these will simply be lost. However, there is one particular result that can be returned, see below.

6.9.1. Expanding your Trajectory via Post-Processing

If your post-processing function expands the trajectory via f_expand() or if your post-processing function returns a dictionary of lists that can be interpreted to expand the trajectory, pypet will start the single runs again and explore the expanded trajectory. Of course, after this expanded exploration, your post-processing function will be called again. Likewise, you could potentially expand again, and after the next expansion post-processing will be executed again (and again, and again, and again, I guess you get it). Thus, you can use post-processing for an adaptive search within your parameter space.

IMPORTANT: All changes you apply to your trajectory, like setting auto-loading or changing fast access, are propagated to the new single runs. So try to undo all changes before finishing the post-processing if you plan to trigger new single runs.

Moreover, your post-processing function can return more than a dictionary, it can return up to five elements.

  1. dictionary for further exploration

2. New args tuple that is passed to subsequent calls to your job function. Potentially these have to be iterables in case you used run_map().

3. New kwargs dictionary that is passed as keyword arguments to subsequent calls to your job function. Potentially these have to be iterables in case you used run_map().

  1. New args for the next call to your post-proc function
  2. New kwargs for the next call to your post-proc function.

6.9.2. Expanding your Trajectory and using Multiprocessing

If you use multiprocessing and you want to adaptively expand your trajectory, it can be a waste of precious time to wait until all runs have finished. Accordingly, you can set the argument immediate_postproc to True when you create your environment. Then your post-processing function is called as soon as pypet runs out of jobs for single runs. Thus, you can expand your trajectory while the last batch of single runs is still being executed.

To emphasize this a bit more and to not be misunderstood: Your post-processing function is not called as soon as a single run finishes and the first result is available but as soon as there are no more single runs available to start new processes. Still, that does not mean you have to wait until all single runs are finished (as for normal post-processing), but you can already add new single runs to the trajectory while the final n runs are still being executed. Where n is determined by the number of cores (ncores) and probably the cap values you have chosen (see Multiprocessing).

pypet will not start a new process for your post-processing. Your post-processing function is executed in the main process (this makes writing actual post-processing functions much easier because you don’t have to wrap your head around dead-locks). Accordingly, post-processing should be rather quick in comparison to your single runs, otherwise post-processing will become the bottleneck in your parallel simulations.

IMPORTANT: If you use immediate post-processing, the results that are passed to your post-processing function are not sorted by their run indices but by finishing time!

6.10. Using an Experiment Pipeline

Your numerical experiments usually work like the following: You add some parameters to your trajectory, you mark a few of these for exploration, and you pass your main function to the environment via run(). Accordingly, this function will be executed with all parameter combinations. Maybe you want some post-processing in the end and that’s about it. However, sometimes even the addition of parameters can be fairly complex. Thus, you want this part under the supervision of an environment, too. For instance, because you have a Sumatra lab-book and adding of parameters should also account as runtime. Thus, to have your entire experiment and not only the exploration of the parameter space managed by pypet you can use the pipeline() function, see also Post-Processing and Pipelining (from the Tutorial).

You have to pass a so called pipeline function to pipeline() that defines your entire experiment. Accordingly, your pipeline function is only allowed to take a single parameter, that is the trajectory container. Next, your pipeline function can fill in some parameters and do some pre-processing. Afterwards your pipeline function needs to return the run function, the corresponding arguments and potentially a post-processing function with arguments. To be more precise your pipeline function needs to return two tuples with at most 3 entries each, for example:

def myjobfunc(traj, extra_arg1, extra_arg2, extra_arg3)
    # do some sophisticated simulation stuff
    solve_p_equals_np(traj, extra_arg1)
    disproof_spock(traj, extra_arg2, extra_arg3)

def mypostproc(traj, postproc_arg1, postproc_arg2, postproc_arg3)
    # do some analysis here

    exploration_dict={'ncards' : [100, 200]}

    if maybe_i_should_explore_more_cards:
        return exploration_dict
        return None

def mypipeline(traj):
    # add some parameters
    traj.f_add_parameter('poker.ncards', 7, comment='Usually we play 7-card-stud')
    # Explore the trajectory
    traj.f_explore({'ncards': range(42)})

    # Finally return the tuples
    args = (myarg1, myarg2) # myargX can be anything form ints to strings to complex objects
    kwargs = {'extra_arg3': myarg3}
    postproc_args = (some_other_arg1,) # Check out the comma here! Important to make it a tuple
    postproc_kwargs = {'postproc_arg2' : some_other_arg2,
                       'postproc_arg3' : some_other_arg3}
    return (myjobfunc, args, kwargs), (mypostproc, postproc_args, postproc_kwargs)

The first entry of the first tuple is you run or top-level execution function, followed by a list or tuple defining the positional arguments and, thirdly, a dictionary defining the keyword arguments. The second tuple has to contain the post-processing function and positional arguments and keyword arguments. If you do not have any positional arguments pass an empty tuple (), if you do not have any keyword arguments pass an empty dictionary {}.

If you do not need postprocessing at all, your pipeline function can simply return the run function followed by the positional and keyword arguments:

def mypipeline(traj):
    return myjobfunc, args, kwargs

6.11. Parameter Optimization

Since pypet offers iterative post-processing and the ability to f_expand() trajectories, you can iteratively explore the parameter space. pypet does not provide built-in parameter optimization methods. However, pypet can be easily combined with frameworks like the evolutionary algorithms toolbox DEAP for adaptive parameter optimization. Using DEAP the evolutionary computation framework shows how you can integrate pypet and DEAP.

6.12. Continuing or Resuming a Crashed Experiment

In order to use this feature you need dill. Careful, dill is rather experimental and still in alpha status!

If all of your data can be handled by dill, you can use the config parameter resumable=True passed to the Environment constructor. This will create a resume directory (name specified by you via resume_folder) and a sub-folder with the name of the trajectory. This folder is your safety net for data loss due to a computer crash. If for whatever reason your day or week-long lasting simulation was interrupted, you can resume it without recomputing already obtained results. Note that this works only if the HDF5 file is not corrupted and for interruptions due to computer crashes, like power failure etc. If your simulations crashed due to errors in your code, there is no way to restore that!

You can resume a crashed trajectory via resume() with the name of the resume folder (not the subfolder) and the name of the trajectory:

env = Environment(continuable=True)


The neat thing here is, that you create a novel environment for the continuation. Accordingly, you can set different environmental settings, like changing the number of cores, etc. You cannot change any HDF5 settings or even change the whole storage service.

When does continuing not work?

Continuing does not work with 'QUEUE' or 'PIPE' wrapping in case of multiprocessing.

Moreover, continuing will not work if your top-level simulation function or the arguments passed to your simulation function are altered between individual runs. For instance, if you use multiprocessing and you want to write computed data into a shared data list (like multiprocessing.Manager().list(), see Sharing Data during Multiprocessing), these changes will be lost and cannot be captured by the resume snapshots.

A work around here would be to not manipulate the arguments but pass these values as results of your top-level simulation function. Everything that is returned by your top-level function will be part of the snapshots and can be reconstructed after a crash.

Continuing might not work if you use post-processing that expands the trajectory. Since you are not limited in how you manipulate the trajectory within your post-processing, there are potentially many side effects that remain undetected by the resume snapshots. You can try to use both together, but there is no guarantee whatsoever that continuing a crashed trajectory and post-processing with expanding will work together.

6.13. Manual Runs

You are not obliged to use a trajectory with an environment. If you still want the distinction between single runs but manually schedule them, take a look at the pypet.utils.decorators.manual_run() decorator. An example of how to use it is given here Starting runs WITHOUT an Environment.

6.14. Combining pypet with an Existing Project

If you already have a rather evolved simulator yourself, there are ways to combine it with pypet instead of starting from scratch. Usually, the only thing you need is a wrapper function that passes parameters from the Trajectory to your simulator and puts your results back into it. Finally, you need some boilerplate like code to create an Environment, add some parameters and exploration, and start the wrapping function instead of your simulation directly. A full fledged example is given here: Wrapping an Existing Project (Cellular Automata Inside!). Or take this script for instance where my_simulator is your original simulation:

from pypet import Environment

def my_simulator(a,b,c):
    # Do some serious stuff and compute a `result`
    result = 42  # What else?
    return result

def my_pypet_wrapper(traj):
    result = my_simulator(traj.a, traj.b, traj.c)
    traj.f_add_result('my_result', result, comment='Result from `my_simulator`')

def main():
    # Boilerplate main code:

    # Create the environment
    env = Environment()
    traj = env.traj

    # Now add the parameters and some exploration
    traj.f_add_parameter('a', 0)
    traj.f_add_parameter('b', 0)
    traj.f_add_parameter('c', 0)
    traj.f_explore({'a': [1,2,3,4,5]})

    # Run your wrapping function instead of your simulator

if __name__ == '__main__':
    # Let's make the python evangelists happy and encapsulate
    # the main function as you always should ;-)